Heatsink & Fan Replacement

[Ralph Wade Phillips]

Howdy!

Does anyone here have any evidence that 70% rubbing alcohol leaves any
residue on anything it cleans? If so, does that residue do anything to
harm heat transfer?

Yes - and no. And, if it's left, yes.

See, that 30% can be quite a few things. Often it's water (which
can be wiped off with a lint free cloth), but if it's "rubbing alcohol" that
30% is quite often various oils - which DON'T wipe of.

So, the answer is, "Read the label."

I've been cleaning things with 70% for a number of decades. My doctor
father cleaned plenty of stuff with it, including wounded humans. I've
never seen any evidence of residue on anything properly cleaned with it.

OK, I haven't been doing microscopic investigations on those surfaces,
but I'm also having a hard time believing that the contact between a
heatsink and a cpu is anywhere near that critical.

And some folks don't believe the world is round, either.

See, how critical it is depends on how fast you need to remove the
heat. Whereas heat sink compounds aren't as good as metal, they beat air
....

They are there to fill in the gaps and imperfections.

I suggest going to a good place, such as Aavid, and studying up on
what HSC is actually SUPPOSED to do ...

Then again, I'm having a hard time understanding why good solid metal to
metal contact isn't enough. That copper heatsink that came with my P4
snaps in very hard. I can't imagine why the contact wouldn't be more
than enough. All that physics in HS and college would seem to support
this -- if I could remember that far back.

Because there's minute imperfections on both surfaces.

Think "Johannsen Blocks" (double check that spelling - I could be
wrong on it!) or "Jo Blocks" as they're quite often called.

I also have a few decades of experience in other areas of life. For
example, a car engine has lots of metal to metal parts that get a heck
of a lot hotter than computers do. True, that have pretty good cooling
systems, but they aren't half as worried about the connection as this
group seems to be.

That's because the motor doesn't fry itself at 140C, but most ICs
start to melt at around that temp ...

Plastic / glass vis metal.

Besides, why didn't Intel say anything about the connection in my P4
boxset? Their instructions just say how to put the chip in and then snap
that heatsink in on top of it. Not a word about cleaning or goop or
anything. Gee, don't they know what they are doing?

It's because they use a better heat spreader, and they have superior
thermal management on the chip. It's not as critical for a P4 as it is for
an Athlon.

Well, some real evidence from someone other than the goop makers would
be real nice.

How about over forty years of experience in semiconductor cooling?
Some small company you might not have heard of such as, oh, Texas
Instruments?

RwP

 

[Guest]

Howdy!



Not no, but HELL, NO!

You use NOTHING that can possibly scratch the HS and/or the CPU.

Use a rubber spreader to burnish the pad off, and either rubbing
alcohol or acetone to remove the rest of the goop.

Bad, BAD advice ...

I'd use only alcohol on a paper towel because according to the
chemical compatiblity database at
www.coleparmer.com/techinfo/chemcomp.asp , acetone severely affects
vinyl (PVC, used for wire insulation, plastic sleeving for
electrolytic capacitors) but not epoxy (circuit boards, chip
packages). I couldn't find anything about polyester (many chip
sockets).

 

[Erez Volach]

"jeffc" <[email protected]> wrote in message

It seems to be much less common than 91%, which every drugstore
carries, but I have an unopened pint of Medalist brand (by McKesson)
99% isopropyl alcohol bought 10 years ago from Drug Emporium for
$1.03.

91% alcohol is perfectly fine for virtually all electronics cleaning
purposes, but 70% sometimes causes corrosion if it remains trapped,
and it takes longer to evaporate. But I wouldn't hesitate to use 70%
for something like removing grease from a CPU and heatsink.

IPA is very hygroscopic. I would venture to say, that a can of 99% IPA would
have to be both very excellent in air-tightness and also have extremely low
water vapor transmission rate, to still be considered 99% after 10 years of
storage - unless it was placed in a dry case with some moisture scavenger
(like silica gel crystals).
once opened, the alcohol will readily absorb moisture from the ambient,
untill reaching (over a considerable span of time) some equilibrium -
depending on air humidity.
91% alcohol is probably more stable moisture-content wise, but i am not
familiar with this concentration. 70% is standard commercial product, and i
am very doubtfull if any medical grade alcohol will have "oil residues".
infact, the medical grade might be ethanol and not IPA.
E.V

 

[Clyde]

Howdy!



Yes - and no. And, if it's left, yes.

See, that 30% can be quite a few things. Often it's water (which
can be wiped off with a lint free cloth), but if it's "rubbing alcohol" that
30% is quite often various oils - which DON'T wipe of.

So, the answer is, "Read the label."


And some folks don't believe the world is round, either.

See, how critical it is depends on how fast you need to remove the
heat. Whereas heat sink compounds aren't as good as metal, they beat air
...

They are there to fill in the gaps and imperfections.

I suggest going to a good place, such as Aavid, and studying up on
what HSC is actually SUPPOSED to do ...


Because there's minute imperfections on both surfaces.

Think "Johannsen Blocks" (double check that spelling - I could be
wrong on it!) or "Jo Blocks" as they're quite often called.


That's because the motor doesn't fry itself at 140C, but most ICs
start to melt at around that temp ...

Plastic / glass vis metal.


It's because they use a better heat spreader, and they have superior
thermal management on the chip. It's not as critical for a P4 as it is for
an Athlon.


How about over forty years of experience in semiconductor cooling?
Some small company you might not have heard of such as, oh, Texas
Instruments?

RwP

OK, what kind of numbers have those 40 years given you? i.e. Does my
not-overclocked P4 3.0GHz with the boxed heatsink and fan run 1%, 5%,
10%, or 25% cooler if I put a good goop in there?

I'm just trying to understand the real world situation here.

Thanks,
Clyde

 

[Clint]

One thing to consider is that you may not be getting solid metal-to-metal
contact. I recently (like last night) pulled off my stock HS from my P4
2.4GHz b processor. Stripped off the thermal pad. Took a look at the base,
and did a light test to see how flat it was (hold it up to the light with a
straight-edge). There was a sizeable dish in the middle of the HS, which
meant that the only contact was on the 2 edges. So I lapped the HS, used
Arctic Silver 3 to put it back together, and saw 6 to 10 degree temperature
drops (idle and load, respectively, in Celsius).

Clint

 

[Clyde]

One thing to consider is that you may not be getting solid metal-to-metal
contact. I recently (like last night) pulled off my stock HS from my P4
2.4GHz b processor. Stripped off the thermal pad. Took a look at the base,
and did a light test to see how flat it was (hold it up to the light with a
straight-edge). There was a sizeable dish in the middle of the HS, which
meant that the only contact was on the 2 edges. So I lapped the HS, used
Arctic Silver 3 to put it back together, and saw 6 to 10 degree temperature
drops (idle and load, respectively, in Celsius).

Clint

Ah, some good numbers. Finally! That's what I've been looking for.
Thanks.

Then again, it might be just your setup. I suppose it's pretty cheap
insurance though.

Clyde

 

[Clint]

It could also be that I screwed up the initial installation on the HS, but I
don't think so. And I'm not sure how representative my HS was of the
general Intel quality. I hope the other ones are, in general, flatter than
mine. I couldn't believe it when I saw the dish in the middle of it. The
surface itself was reasonably smooth, but there was no way it was getting
close contact with the CPU.

Clint